Treffer: Uncertainty-aware physics-data fusion for RUL prediction for manufacturing tools.

Accurate prediction of tools’ remaining useful life (RUL) is essential for maintaining optimal performance of the associated manufacturing system. Effective RUL prediction requires understanding degradation mechanisms and quantifying their inherent uncertainty, motivating the fused approaches that integrate physics-based and data-driven models. This study proposes a hybrid physics-based data-driven framework to predict tool RUL with interpretability, uncertainty, and flexibility. This framework aligns with the definition of tool capability and creates bidirectional interfaces among different models. First, physics-based degradation models are deployed based on known processing and degradation mechanisms to simulate primary degradation data. Then, data-driven techniques identify dominant degradation patterns and model baseline trend and associated uncertainty. These data-driven insights are further projected to the physical model, allowing the modelled patterns to be interpreted within the context of the actual degradation procedure. This integration supports more reliable RUL predictions by combining the strengths of both approaches, i.e. physical interpretability and statistical uncertainty. Compared with conventional methods, the proposed model outperforms in fidelity, accuracy, and adaptability to changes in product quality requirements. Its improved performance is attributed to three factors: incorporation of physics data distribution, quantification of uncertainty during processing, and the responsiveness to real-time updates in processing requirements. [ABSTRACT FROM AUTHOR]

Copyright of International Journal of Production Research is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)